The present review was based on the available literature to date, which consists predominantly of case reports and small case series, all graded as having low-quality evidence. Table 1 and Table 2 contain a partial list of case reports,[4,9–31] corresponding only to the descriptions of single cases with detailed information. Such reports refer to patients with laboratory-confirmed COVID-19. In only one case, although multiple nucleic acid tests were negative, the infection was diagnosed on the basis of both chest computerized tomography (CT) scan findings and close contact with people with COVID-19. Case series with more than three patients are not contained in these tables and are described in depth throughout the text. The only report with higher-quality evidence is that by Galván Casas et al. on a large cohort of patients, which allowed the authors to stratify COVID-19-related skin manifestations into five categories. Overall, the case reports suggest that skin lesions developed more often after the onset of COVID-19 symptoms, with a variable latency period. A simultaneous onset was sometimes noted, especially for fever. More rarely, skin lesions occurred in the prodromal phase, shortly before the appearance of typical COVID-19 symptoms.[9,27] Reports concerning asymptomatic patients also exist. Different clinical patterns were described.
Classification of COVID-19-associated Cutaneous Manifestations
A classification of cutaneous manifestations of COVID-19 has been proposed based on the results of a prospective nationwide consensus study in Spain using a representative sample of 375 cases. Five clinical patterns were recognized: (i) acral areas of erythema with vesicles or pustules (pseudo-chilblains) (19%); (ii) other vesicular eruptions (9%); (iii) urticarial lesions (19%); (iv) maculopapular eruptions (47%); and (v) livedo or necrosis (6%). Vesicular eruptions were found to appear early in the course of the disease (before other symptoms in 15% of cases), chilblain-like lesions frequently appeared late over the disease course, and the remaining patterns tended to develop during the illness phase. Chilblain-like lesions also tended to have a longer duration than the other forms. A gradient of severity of COVID-19 could be observed ranging from less severe disease in acral lesions to most severe in the case of livedoid presentations. Nevertheless, transient livedoid eruptions have also been reported in the literature.
Based on our experience, our review of the literature and the classification by Galván Casas et al., who studied a cohort of 375 Spanish patients, we subdivided the reported cutaneous lesions into six main clinical patterns: (i) urticarial rash (Figure 1a),[6,8–16] (ii) confluent erythematous–maculopapular–morbilliform rash (Figure 1b),[3,6,14,15,17–19,21–27] (iii) papulovesicular exanthem (Figure 1d),[6,20,33–35] (iv) chilblain-like acral pattern (Figure 1c),[8,29,30,33,36–41] (v) livedo reticularis–livedo racemosa-like pattern (Figure 1f)[28,31] and (vi) purpuric 'vasculitic' pattern (Figure 1e).[31,33] Although livedo reticularis/racemosa and purpura could be included in the same group, we decided to split them into two separate categories, keeping in mind the hypothesis that the former has a vasculopathic origin while the latter has a true vasculitic pathogenesis.
COVID-19-associated cutaneous manifestations. (a) Urticarial rash. (b) Combination of confluent erythematous rash on the chest with petechial lesions on the abdomen and upper extremities. (c) Acral chilblain-like lesions on the foot. (d) Vesicular exanthem. (e) Palpable purpura on the knees. (f) Livedo racemosa-like lesions on the thighs. All of the photographs belong to the authors' own collection.
For clarity of exposition, the six patterns were assigned to two broader categories: inflammatory or exanthematous eruptions, including urticarial rash, confluent erythematous–maculopapular–morbilliform rash and papulovesicular exanthem (Table 1); and vasculopathic or vasculitic lesions, including chilblain-like acral pattern, livedo reticularis–livedo racemosa-like pattern and purpuric 'vasculitic' pattern (Table 2).
Exanthems and Other Inflammatory Eruptions
These encompass confluent erythematous, maculopapular, morbilliform and urticarial presentations, similar to nonspecific rashes in the course of common viral infections. In the large cohort of 375 patients reported by Galván Casas et al., exanthems were the most commonly reported skin manifestations. In particular, maculopapular eruptions, urticarial lesions and vesicular eruptions accounted for 47%, 19% and 9% of all cutaneous manifestations, respectively. The main differential diagnoses in such presentations are cutaneous drug reactions. Itch was of variable intensity and sometimes absent. Urticarial and maculopapular exanthems have been suggested to appear simultaneously with systemic symptoms, to last approximately 1 week and to be associated with severe COVID-19. The possible presence of urticarial manifestations is unsurprising given that viral agents have been recognized as an important trigger of acute urticaria and urticaria-like rash, especially in children.[42,43] Erythematous flexural lesions were also noted, as well as localized plaques on the heels, and a papulosquamous eruption, clinically reminiscent of pityriasis rosea. Gisondi et al. gave a brief mention of a diffuse papular eruption seen in a woman with COVID-19 febrile infection. Lesions resembling erythema multiforme have also been described.
In some case reports (Table 1) the outcomes were not specified. When available, the data on the natural evolution generally indicated an overall short duration of skin involvement.
Hedou et al. analysed, in a prospective study, the incidence and types of cutaneous manifestations associated with COVID-19 in France. In total, 103 patients with confirmed infection (71 women) with a mean age of 47 years were evaluated. Among them, 76 were treated at home, while 23 were admitted to conventional hospital wards and four to intensive care units. Only five (5%) presented with skin manifestations: erythematous rash (n = 2) and urticaria (n = 2), mainly located on the face and upper body, and one case of oral herpes simplex virus reactivation in an intubated patient. Cutaneous signs appeared during the prodromal phase in a patient with urticaria and during the illness in the remaining patients. All eruptions were associated with itching and disappeared within 6 days (median 2 days).
A further retrospective observational study recorded the occurrence of inflammatory lesions, usually a few days after the onset of COVID-19 systemic symptoms, in seven patients, including exanthem (n = 4), varicella-like vesicles (n = 2) and cold urticaria (n = 1). An eruption worthy of separate comment is that characterized by vesicular or papulovesicular lesions resembling varicella (Table 1).[6,20,34,35] In the cohort studied by Galván Casas et al., vesicular lesions usually appeared in middle-aged patients, before the onset of systemic symptoms in 15% of cases, and were associated with intermediate severity of COVID-19.
Eight Italian dermatology units collected clinical data on 22 patients with COVID-19 (16 male, median age 60 years) and varicella-like lesions. In that study, which two authors of the present article (A.V.M. and G.G.) contributed to, the skin lesions usually appeared 3 days after systemic symptoms and disappeared after 8 days, without scarring. Systemic symptoms were most commonly fever (98%), followed by cough (73%), headache (50%), weakness (50%), coryza (46%) and dyspnoea (41%), and less frequently hyposmia, hypogeusia, pharyngodynia, diarrhoea and myalgia. Death occurred in three patients. The authors reported that typical characteristics of the rash are frequent trunk involvement, scattered distribution and mild or absent pruritus, speculating that this kind of eruption can be an almost specific COVID-19-associated cutaneous manifestation. The case of a paediatric patient included in this series has been described in detail in another subsequently published report.
Tammaro et al. presented data from their combined experience in Rome, Italy, and Barcelona, Spain, and focused on lesions similar to those found in infections caused by members of the family Herpesviridae. In particular, they described the case of a woman from Barcelona who developed numerous, isolated vesicular lesions on the back 8 days after the diagnosis of COVID-19. In the same paper, they identified, among 130 patients with COVID-19 hospitalized in Rome, two (1·5%) cases presenting with isolated erythematous–vesicular herpetiform lesions on their trunk accompanied by mild pruritus.
Some cutaneous rashes encountered in association with COVID-19 were characterized by a petechial component,[15,19,24,25] which may have been secondary to thrombocytopenia. These cases were not included in the purpuric 'vasculitic' pattern of our classification because, in our opinion, petechial lesions in such circumstances were only an accompanying finding of inflammatory nonvasculitic eruptions, which were predominantly characterized by macular or maculopapular lesions. Petechial lesions could also represent a secondary phenomenon during the natural evolution of the exanthem, as described in a case in which a maculopapular rash became purpuric 1 week later. Erythematous–purpuric, maculopapular and petechial lesions with a tendency towards a flexural or periflexural distribution have also been reported.[24,25]
Vasculopathic or Vasculitic Lesions
In a retrospective observational French study, seven patients presented with vascular lesions as follows: violaceous macules with 'porcelain-like' appearance (n = 1), livedo (n = 1), non-necrotic purpura (n = 1), necrotic purpura (n = 1), eruptive cherry angiomas (n = 1), chilblain-like lesions alone (n = 1) and chilblain-like lesions associated with Raynaud phenomenon (n = 1). Manifestations defined as petechiae, tiny bruises or livedoid eruptions have been linked to COVID-19, as reported by an article in a news magazine published by the Society of Hospital Medicine. Furthermore, it was speculated that these manifestations might be a result of small-blood-vessel occlusion, whose pathogenic mechanisms (e.g. neurogenic, microthrombotic or immune complex mediated) are as yet unknown.
Intriguingly, Magro et al. documented the presence of thrombotic microvascular damage in the lung and/or skin of five patients with critical COVID-19. In three of these patients, purpuric or livedoid skin lesions (Table 2) characterized by a pauci-inflammatory microthrombotic vasculopathy were described (Table 3). Deposition of complement components within the skin and lung microvasculature was detected (Table 3), with colocalization of SARS-CoV-2-specific spike glycoproteins. It was suggested that severe disease might induce a catastrophic generalized microvascular injury syndrome mediated by intense activation of the alternative and lectin complement pathways and an associated procoagulant state.
It is worth mentioning that a group of Chinese authors reported the development of severe acro-ischaemia in a few patients in intensive care units in Wuhan, manifesting as finger or toe cyanosis, skin blisters and dry gangrene, resulting from a hypercoagulable status or confirmed disseminated intravascular coagulation.
Alramthan et al. hypothesized that acral ischaemia could be responsible for the acral red-purple perniosis-like papules noted in two asymptomatic young women. Very mild processes of intravascular coagulation and microthrombosis have been implicated in the pathogenesis of transient localized livedo reticularis-like lesions. Biopsies were not taken in such cases because of the fleeting nature of the skin lesions. Therefore, livedoid skin lesions might be mediated by an occlusive thrombotic microvasculopathy of variable magnitude. In many cases, these lesions are likely to be mild and transient. Nevertheless, Manalo et al., describing cases of transitory livedo reticularis-like patches (Table 2), declared that platelet count, coagulation studies, and assessment of fibrin degradation products in these patients would be enlightening. In patients with severe COVID-19, vasculopathic lesions might be more pronounced and/or persistent, being the epiphenomenon of uncontrolled systemic microvascular damage or coagulopathy. Interestingly, livedoid and necrotic lesions have been suggested to occur in elderly patients with severe systemic symptoms, likely representing the cutaneous manifestations associated with the highest rate of COVID-19-associated mortality.
In our classification, we distinguished a specific purpuric pattern that was defined as 'vasculitic' in order to differentiate it from the petechial component of some exanthematous eruptions, as well as from livedo related to occlusive or microthrombotic vasculopathy. This pattern is likely to be extremely rare.[31,33] In our experience, we observed a case of palpable purpura on the knees (Figure 1e) and another patient with purpuric and necrotic lesions of the lower legs, clinically resembling cutaneous leucocytoclastic vasculitis (unpublished data). In such cases, a vasculitic aetiology might be implicated, although histopathological data were unavailable and further studies are required. Interestingly, among the three vasculopathic lesions in the setting of severe COVID-19 documented by Magro et al., the retiform purpura also showed prominent leucocytoclasia, suggesting the coexistence of a vasculitic process (Table 3).
A peculiar cutaneous manifestation that is raising particular interest among dermatologists is represented by acral lesions resembling chilblains.[8,29,30,33,36–41] These atypical presentations occurred in young patients confined at home during springtime, in the absence of cold exposure, comorbidities or other potential triggers. The rapid outbreak of perniosis-like lesions, in parallel with the COVID-19 pandemic and reflected in the ever-growing number of reports of similar cases from various countries, has led to the hypothesis of a relationship between the disease and these clinical observations, although more data are needed to draw definite conclusions. In many cases, laboratory confirmation of the infection was not performed or gave negative results, and/or complete information about testing was not available. Nevertheless, various reports have suggested that such acro-ischaemic chilblain-like lesions predominantly affect children and young people, with an occasional history of systemic symptoms preceding cutaneous lesions.[37,39,41]
The acral eruption was frequently characterized by erythematous–violaceous papules and macules, with possible bullous evolution, or digital swelling. Feet were more frequently affected than hands. Nonacral sites were rarely involved and sometimes showed lesions with a targetoid appearance.[39,41] In some cases, this was likely to represent true virus-induced erythema multiforme, theoretically triggered by SARS-CoV-2 (personal observation).
In the study by Galván Casas et al., chilblain-like lesions more often affected younger patients, lasted for approximately 13 days, occurred later in the course of COVID-19 and were associated with less severe disease.
In their series of 14 patients with acral perniosis-like lesions, Recalcati et al. found no abnormalities of laboratory parameters (complete blood count, C-reactive protein, lactate dehydrogenase and D-dimer; serology for Epstein–Barr virus, cytomegalovirus, coxsackie and parvovirus B19). RT-PCR test for SARS-CoV-2 from a swab was performed in very few cases, with negative results.
Fernandez-Nieto et al. retrospectively reviewed the data of 132 patients with acral lesions (mean age 19·9 years) between 5 March and 15 April 2020. Two main clinical patterns of acute acro-ischaemic lesions were recognized, resembling chilblains and erythema multiforme, respectively. Of these patients, 41% had close contact with people with COVID-19 and 14·4% were clinically diagnosed with COVID-19 (16 patients had skin lesions after a mean period of 9·2 days from the onset of COVID-19 symptoms, and three patients simultaneously). None of the patients had pneumonia or any other complication. In 11 patients, an RT-PCR assay was performed from a nasopharyngeal swab, after the appearance of skin lesions, and yielded positive results in only two patients.
It has been speculated that chilblain-like lesions could represent late manifestations of COVID-19 in young healthy patients.[37,41] Future studies with serological examinations are necessary to corroborate this assumption. The lesions might result from a delayed immunological reaction against viral particles in asymptomatic or paucisymptomatic forms of COVID-19.[33,41] An increased type I interferon reaction has also been implicated.[30,33]
Histopathological Studies and Other Laboratory Assessments of Skin Samples
Histopathological examination of lesional skin was performed in a limited number of cases (Table 3). Inflammatory lesions generally showed nonspecific features.[10,15,16,22,25–27,34] The histopathological picture of exanthematous eruptions was usually compatible with viral exanthem. The clinicopathological features found by Gianotti et al. were reminiscent of Grover disease in one patient and consistent with viral exanthem in the other two. In the latter two cases peculiar findings were noted, with early dermal microthromboses in one and signs suggestive of lymphocytic vasculitis in the other. Targetoid erythema multiforme-like lesions of the elbows, observed in patients with concomitant acral perniosis-like lesions, were characterized only by a mild superficial perivascular dermatitis.
Chilblain-like lesions were characterized by lymphoid–lymphoplasmacytic infiltrate in the dermis, possibly extending to the hypodermis, and by a prevalent perivascular pattern.[30,41] Signs of endothelial activation or plump endothelial cells in the venules surrounded by infiltrate were detected.[30,41]
The purpuric–livedoid skin lesions observed by Magro et al. in three critical patients with COVID-19 showed signs of thrombotic microvascular damage. In addition, histopathological examination of the livedoid lesions revealed no signs of inflammation in one case and mild perivascular lymphocytic infiltration in another. In contrast, the retiform purpura of a third patient showed neutrophilia with prominent leucocytoclasia along with necrotic changes in the epidermis and adnexa. Immunohistochemical assessment demonstrated significant deposition of complement components within the skin microvasculature (Table 3). In two cases, a biopsy taken from the normal-appearing skin showed microvascular deposits of C5b-9 throughout the dermis.
Direct immunofluorescence studies were performed only in a case of erythematous–oedematous figurate plaques and in a patient with chilblain-like acral lesions, providing negative results in both cases.
The British Journal of Dermatology. 2020;183(3):431-442. © 2020 Blackwell Publishing